The AI would build a proof of correctness, which would be then verified in a proof checker (not AI).
And how do you prove that the proof of correctness is not just a proof that 1=1? LLMs "cheating" on things is rather common.
alt Hacker News
And how do you prove that the proof of correctness is not just a proof that 1=1? LLMs "cheating" on things is rather common.